Update evaluation.py

evaluation.py

@@ -113,7 +113,8 @@ def retrieve_ground_truths(question, dataset):
                     "context_relevance": instance['relevance_score'],
                     "context_utilization": instance['utilization_score'],
                     "completeness": instance['completeness_score'],
-                    "adherence": adherence_numerical
+                    "adherence": adherence_numerical,
+                    "response_time": time_taken
                 }
                 return instance_response, ground_truth_metrics  # Return the ground truth response immediately
     return None,None  # Return None if no match is found
@@ -171,7 +172,8 @@ def calculate_metrics(question, q_dataset, response, docs, time_taken):
         "context_relevance": context_relevance(question, docs),
         "context_utilization": context_utilization(response, docs),
         "completeness": compute_cosine_similarity(response, ground_truth_answer), #completeness(response, ground_truth_answer),
-        "adherence": adherence(response, docs)
+        "adherence": adherence(response, docs),
+        "response_time": time_taken
     }
 
     store_rmse(question, predicted_metrics, ground_truth_metrics)